INDEX

Explanations

predicting prices, models, types

The neuron fires on unusually long or low‐frequency subword tokens—i.e. rare or “exotic” pieces of words.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

өп

0.92

 halen

0.78

سن

0.73

 Demp

0.73

abhavena

0.73

 zaken

0.73

 poderá

0.72

 rheumatism

0.72

න්ද

0.71

 säga

0.71

POSITIVE LOGITS

↵↵

0.72

 Tiempo

0.72

LOCK

0.66

ҙ

0.63

to

0.61

tiempo

0.61

 다시

0.60

zw

0.60

вла

0.59

 enabled

0.59

Activations Density 0.000%